4 Cloud Data Practices Every Healthcare Organization Should Be Using

4 Cloud Data Practices Every Healthcare Org Should Be Using
Executive Summary

In this webinar, Skip Marsh, Principal Solution Architect at Silk, and the host discuss the four essential cloud data practices healthcare organizations need to manage massive data growth, control cloud costs, and enable innovation.
Drawing from customer examples like Sentara Health and Franciscan Health, they explore how healthcare data is expanding exponentially — doubling every 73 days — and how advanced storage strategies, copy data management, and Silk’s software-defined storage platform can help organizations manage that growth while accelerating analytics and AI initiatives.

Key Takeaways

📈 Healthcare Data Growth Is Exploding: Data volumes double every 73 days due to EHRs, imaging, wearables, and AI workloads.

💸 Cloud Costs Are Surging: The shift to cloud introduces unexpected expenses; controlling data sprawl and copies is key.

⚙️ Copy Data Management Is Critical: Efficient cloning and snapshot strategies reduce latency, control cost, and improve agility.

🤖 AI Needs Safe Access to Data: Enabling AI and analytics requires rapid, governed copies — not direct access to production systems.

🚀 Silk Bends the Cost and Time Curve: Through thin clones, instant snapshots, and data reduction, Silk dramatically reduces refresh time and storage spend for healthcare environments.

[00:00:00] Introduction

Host:
Welcome, everyone, and thanks for joining today’s webcast. I’m here with Skip Marsh, Principal Solution Architect at Silk.
Today we’re diving into four cloud data practices every healthcare organization should be using.

This discussion is grounded in real-world experience with customers like Sentara Health and Franciscan Health, focusing on how healthcare data growth is transforming cloud storage and cost management.

[00:01:00] The Data Explosion in Healthcare

Host:
Healthcare data is doubling every 73 days. That growth comes from several key drivers — imaging, genomic data, expanding EHR systems, wearables, and the rise of AI and analytics.

We’re seeing two main types of healthcare data:

Unstructured data – documents, notes, emails, videos, and audio.

Structured data – patient demographics, lab results, and clinical metrics stored in relational databases.

Structured data is especially valuable for analytics and AI. Managing both efficiently — while ensuring accessibility and compliance — is now a top challenge.

[00:03:00] Practice #1: Managing Cloud Cost Growth

Host:
Nearly every healthcare customer we talk to says the same thing: they’re surprised by their cloud bill.

Skip Marsh (Silk):
Absolutely. Cloud costs in healthcare are skyrocketing. The challenge isn’t just the size of the data — it’s how it’s being managed.
Many organizations create multiple copies of data sets, and each one consumes storage, compute, and budget.

Deleting data isn’t an option in healthcare because of compliance and retention requirements, so the key is smarter data management, not deletion.

[00:04:00] Practice #2: Enabling Data Access Across Teams

Host:
Within healthcare, you’ve got payers, providers, analytics, AI, and reporting teams — all needing access to the same data. That’s a recipe for ballooning copies and complexity.

Skip:
Right. Every team — from data scientists to analysts — needs access to EHR or claims data. But they can’t all touch production systems.

The top question we get is: “How do we get more data copies — fast — without risking downtime or massive cloud bills?”
The answer lies in how you replicate and distribute those datasets.

[00:06:00] Practice #3: Implementing Copy Data Management

Host:
Let’s talk about copy data management — one of the biggest opportunities for efficiency.

Skip:
Traditional approaches rely on database-level copies — slow, resource-heavy, and disruptive.
Some healthcare orgs can only refresh environments once or twice a year because it’s so complex.

With Silk, we use instant snapshots and thin clones, so any environment — even 50–100 TB databases — can be copied and refreshed in minutes instead of days.
That eliminates latency between production and analytics environments and saves huge amounts of money.

[00:08:00] Maintenance Windows and Production Impact

Skip:
Many hospitals run lengthy ETL (extract, transform, load) jobs during maintenance windows.
As data grows, those windows get longer — and often spill into business hours. Doctors and analysts expect access by 7 a.m., but ETL jobs can still be running.

Using instant cloning, those windows shrink from hours to minutes, ensuring clinicians always have access to live data when they start their day.

[00:10:00] Practice #4: Supporting AI and Advanced Analytics

Host:
Let’s talk about AI. It’s a growing focus in healthcare — for patient outcomes, diagnostics, and efficiency.

Skip:
Exactly. Every organization has an AI initiative now, but they can’t give AI systems unrestricted access to EHR data.
AI models (especially agentic AI) can generate unpredictable workloads, so direct access to production databases is unsafe.

Silk enables safe, rapid access through fresh clones of large EHR systems — letting AI teams experiment at high speed without impacting live systems.

[00:12:00] Compounding Complexity

Host:
So to summarize:

Massive data volumes

More teams needing access

Copy data challenges

AI adding new workloads

That’s a perfect storm — and the healthcare industry is feeling it first.

Skip:
Right. Healthcare data isn’t just growing — it’s compounding. And AI doesn’t just read data; it generates new data, which must also be stored and governed.

[00:13:00] Silk’s Solution: Software-Defined Cloud Storage

Skip:
Silk’s platform is a software-defined cloud storage layer designed for performance and scalability.

We can fine-tune performance in real time, scale compute and storage independently, and eliminate the need for massive VMs just to meet throughput demands.
Customers achieve smaller VM footprints, cutting infrastructure and licensing costs dramatically.

[00:14:00] Instant Copying and Cloning

Skip:
We’ve patented several technologies for instant copy data management.
With Silk, you can take a snapshot of any size database — even 100 TB — and create thin clones mounted to any environment in under two minutes.

These clones are space-efficient and perfect for dev/test, analytics, or AI workloads.
Customers can run multiple clones in parallel with zero impact on production and at a fraction of the cost of native cloud tools.

[00:16:00] Performance and Cost Savings

Skip:
Silk helps organizations “bend the cost curve” — and the time curve.
Customers report:

5–10× faster performance on I/O-intensive workloads

ETL windows reduced from 7–10 hours down to 15 minutes

Cloud savings of up to $22 million over three years

These gains come from data reduction (compression and deduplication) and the ability to use smaller VM shapes without losing throughput.

[00:20:00] Customer Example: Epic Clarity Database

Skip:
Here’s an example from a major Epic customer.
Their Clarity database (around 45 TB) used to be offline 7–10 hours per night for ETL jobs.
Using Silk snapshots and thin clones, they reduced downtime to 15 minutes per night, keeping the system online 23 hours and 45 minutes daily.

They now create three clones nightly for analytics, backups, and utility tasks — all running independently without impacting production performance.

[00:23:00] Integrating Data Masking and Compliance

Skip:
For payer systems (like QNXT and Facets), we integrate with data masking tools to protect PHI and PII.
Organizations can now give offshore development teams secure, masked data copies that refresh in 15 minutes — instead of taking 5–7 days.

This not only accelerates development but also maintains compliance with HIPAA and other data protection standards.

[00:27:00] Real-World Results

Skip:
In one case, a customer saved 600 TB of physical cloud storage through Silk’s compression and cloning.
Another achieved 13:1 data reduction, provisioning 1.7 PB of logical storage while using only 135 TB of physical space — a massive cost savings.

[00:32:00] High-Performance Benchmark

Skip:
In a performance test with an online trading customer using Azure, Silk achieved 34 GB/s throughput with sub-millisecond latency — results even Azure engineers found remarkable.

This shows how Silk not only supports healthcare workloads but also scales to enterprise-grade, real-time data performance for any industry.

[00:34:00] Closing Remarks

Host:
Thanks, Skip, for sharing these insights.
For everyone watching — you can find resources and case studies on Silk’s website, including details about Sentara Health’s success and Silk’s Business Value Assessments, which quantify cost and time savings.

If you’d like a customized demo or to explore how Silk can support your Epic, Facets, or AI initiatives, contact us through our site or via the links below.

Thank you for joining us — and have a great day!